Evaluation on soil water validity using optimum partitioning clustering method

For promoting the insufficiency in evaluating soil water availability (SWA) through soil water characteristic curve, soil water characteristic curves were partitioned by means of optimum partitioning clustering theory and clustering centre values were selected as evaluation indices. A projection pursuit evaluation model was constructed to evaluate soil water validity of 33 soil samples taken from 11 main tobacco-growing areas in Guizhou Province, China. The results were compared with those of conventional method by using soil physical properties as evaluation indices. The results showed that the selected indices could reflect soil water regime of tobacco field, and the results of the evaluation according to the indices were more comprehensive than those from conventional method.